You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Yun Tang (Jira)" <ji...@apache.org> on 2023/03/31 03:17:00 UTC

[jira] [Commented] (FLINK-31573) Nexmark performance drops in 1.17 compared to 1.13

    [ https://issues.apache.org/jira/browse/FLINK-31573?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17707095#comment-17707095 ] 

Yun Tang commented on FLINK-31573:
----------------------------------

I think https://github.com/nexmark/nexmark/pull/43 could help mitigate this problem.

> Nexmark performance drops in 1.17 compared to 1.13
> --------------------------------------------------
>
>                 Key: FLINK-31573
>                 URL: https://issues.apache.org/jira/browse/FLINK-31573
>             Project: Flink
>          Issue Type: Bug
>          Components: Runtime / Checkpointing
>    Affects Versions: 1.17.0
>            Reporter: Qingsheng Ren
>            Priority: Critical
>
> The case was originally [reported|https://lists.apache.org/thread/n4x2xgcr2kb8vcbby7c6gwb22n0thwhz] in the voting thread of 1.17.0 RC3. 
> Compared to Flink 1.13, the performance of Nexmark in 1.17.0 RC3 drops ~8% in query 18. Some details could be found in the [mailing list|https://lists.apache.org/thread/n4x2xgcr2kb8vcbby7c6gwb22n0thwhz]. 
> A further investigation showed that with configuration {{execution.checkpointing.checkpoints-after-tasks-finish.enabled}} set to false, the performance of 1.17 is better than 1.16. 
> A fully comparison of Nexmark result between 1.16 and 1.17 is ongoing.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)